Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events
(Huete)
Scaling photosynthesis in tropical systems: from
forest to savanna, from seasons to extreme events
Proposal Submitted in Response to NASA ROSES Terrestrial Ecology Program
NRA: NNH09ZDA001N-TE
Team Members
PI: Dr. Alfredo R. Huete
Dept. Soil, Water & Environmental Sciences
University of Arizona, Tucson, AZ USA
ahuete@email.arizona.edu
Co-Investigator: Dr. Dennis Dye
U.S. Geological Survey, Flagstaff, AZ USA
ddye@usgs.gov
Co-Investigator: Dr. Scott R. Saleska
Dept. Ecology and Evolutionary Biology
University of Arizona, Tucson, AZ USA
saleska@email.arizona.edu
Co-Investigator: Dr. Yosio E. Shimabukuro
Instituto Nacional de Pesquisas Espaciais (INPE), São José dos Campos, Brazil
yosio@dsr.inpe.br
Co-Investigator: Dr. Humberto da Rocha
Dept de Ciências Atmosféricas, Universidade de São Paulo, São Paulo, Brazil
humberto@model.iag.usp.br
Co-Investigator: Dr. Antonio Manzi
National Institute for Amazonia Research (INPA), Manaus, Amazonas, Brazil
manzi@inpa.gov.br
Collaborator: Dr. Hideki Kobayashi
Department of Environmental Science, Policy, and Management,
University of California, Berkeley, Berkeley, CA, USA
hkoba@nature.berkeley.edu
Project Period: January 1, 2010 to December 31, 2012
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
Proposal Summary: A fundamental unanswered question for global ecology is the vulnerability
of tropical forests to climate change, up to and including widespread Amazon forest collapse and
conversion to savanna due to global warming-induced drought, a projection of some coupled
carbon/climate models. Remote sensing methods, by observing broadscale vegetation responses
to climatic variability, offer potentially powerful insight into this question. For example, recent
studies using MODIS remote sensing products have detected positive vegetation responses to
seasonal drying and interannual drought which if accurate, suggest that tropical vegetation is
more robust than most ecosystem models suggest, with implications for long-term vulnerability
to climate change.
TECHNICAL CHALLENGE: However, multiple satellite products collected from the AVHRR
and MODIS platforms show different, inconsistent seasonal and interannual patterns over
Amazon tropical rainforests with some sensor products showing canopy drying and negative
forest responses to dry periods. Some argue that the observed positive responses are an artifact of
aerosol contamination. Plot-based studies reporting increased tree mortality during the 2005
Amazon drought have led some to conclude that there is an inconsistency between ground plots
(showing long-term mortality) and remote sensing (showing short-term green-up), though the
two observations may be reconciled once the different timescales and different locations of the
different effects are taken into account. More fully validated observations are needed to build
confidence in remotely sensed ecosystem function before community acceptance of observed
results, and to realize the full potential of the future array of space-based spectral sensors,
including the VIIRS mission and hyperspectral HyspIRI mission proposed as part of the NRCs
Decadal Survey.
OBJECTIVES AND METHODS: We propose a study to investigate the mechanisms underlying
the divergent responses to the 2005 Amazonian drought, and to decisively test competing
hypotheses about mechanisms underlying green-up forest response to dry periods (including:
increased light,, e.g. increased LUE under aerosol-induced increases in diffuse radiation fraction,
evolutionarily prescribed leaf-flush phenology, and the null hypothesis of contamination
artifacts). We propose three components. FIRST, we will compare independent measures of the
seasonality of: (1) vegetation photosynthetic capacity (at both individual leaf and ecosystem
scales), (2) atmospheric characteristics (aerosol optical depth and cloud cover), (3) surface
radiation components (including direct and diffuse radiation), and (4) vegetation spectral
reflectances (and indices) from satellites and from in-situ AERONET sunphotometers, at four
sites across the Amazon, chosen to span a range of tropical ecosystems and seasonal atmospheric
aerosol characteristics. SECOND, we will use hyperspectral observations at both the tower-level
and from space (via Hyperion imagery) to diagnose key physiological processes and further test
the accuracy and interpretation of the high-frequency moderate-resolution MODIS products.
THIRD, we will use the first airborne LIDAR dataset obtained for the Amazon to parameterize a
sophisticated 3-D canopy photosynthesis model (FLiES) in order to scale up our integrated
understanding of vegetation characteristics (including leaf spectral reflectance), radiation
components (including aerosol-, cloud-, and subcanopy-influenced effects of diffuse radiation
fraction and angular distribution), and the seasonality thereof.
SCIENTIFIC SIGNIFICANCE AND RELEVANCE TO NASA GOALS. This work will give
new insight into the fundamental question of tropical forest "vulnerability to global
climate change" (Subelement 1 of the ROSES call); will produce a fusion of
hyperspectral with high-frequency moderate resolution datasets, thereby providing a
focused example of how HyspIRI Decadal Survey mission data may be used "to
2
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
understand seasonal expressions and cycles" (Subelement 2); and will synthesize and
integrate substantial amounts of data from LBA, a "past NASA Terrestrial Ecology
research project" (Subelement 3).
Table of Contents
Cover Page ........................................................................................................................
Page
1
Summary ..........................................................................................................................
2
1. Scientific/ Technical Management ..............................................................................
1.1 Rationale and Background .........................................................................................
1.2 Example: 2005 Drought ............................................................................................
1.3 Objectives ...................................................................................................................
2. Technical approach and methodology ...........................................................................
2.1 Task 1 Assessment of seasonal expressions .................................................................
2.2 Task 2 Analysis of environmental drivers ....................................................................
2.3 Task 3 Scaling seasonal leaf phenology ........................................................................
2.4 Data acquisition from four tower sites .........................................................................
2.5 Measurements ...............................................................................................................
2.6 FLiES model (3-d canopy model) .................................................................................
2.7 Seasonal patterns of solar radiation ..............................................................................
3. Management Plan ..........................................................................................................
3.1 Management structure and collaboration ..................................................................
3.2 Milestones and timeline .............................................................................................
3.3 Expected outcomes and products ..............................................................................
3.4 Facilities ....................................................................................................................
4. References ...................................................................................................................
5
5
7
8
9
10
11
12
14
15
16
17
18
18
18
19
19
20
Brazil Tower Site Facilities ...........................................................................................
3
23
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
1. Scientific/Technical/Management
1.1 Rationale and Background: A fundamental unanswered question for global ecology is the
degree to which tropical forests are vulnerable to climate change, up to and including widespread
Amazon forest collapse and conversion to savanna due to global warming-induced drought, as
predicted by some coupled carbon/climate models (Betts et al., 2004). Many climate change
scenarios projected for the 21st century predict the Amazon to become drier due to stronger El
Niño conditions and an enhanced North-South Atlantic sea surface temperature gradient (Cox et
al. 2008). There is increasing concern on savannization of the Amazon and the consequent
release of large stocks of carbon to the atmosphere associated with forest dieback and the
conversion of tropical forest to savanna-like ecosystems (Sitch et al., 2008; Malhi et al., 2009).
The vulnerability of tropical forest systems to climate change depends not just on the physical
climate system, but on biological response of forests to initial climate changes. While much
attention has focused on the Amazon die-back scenario (Betts et al., 2004), other models imply
forest persistence (Friedlingstein et al., 2006). Differing modeled fates of the forest were due to
differences among models in representation of forest function, not just differences in
representations of climate (Sitch et al., 2008). Current knowledge is insufficient to determine
which model representation of vegetation function are most consistent with real forest
ecosystems, but continuing observations from satellites and from the network of eddy flux
towers provide tools that can rigorously test mechanisms of forest-climate interactions, and
hence, provide critical insight into the potential future of tropical forests.
For example, recent studies using MODIS remote sensing products have detected positive
‘greening’ vegetation responses to seasonal drying (Huete et al., 2006; Myneni et al., 2007) and
interannual drought (Saleska et al., 2007) which if accurate, suggest that tropical vegetation, by
apparently responding to increased dry-period light availability, is more robust than most
ecosystem models suggest, with implications for long-term vulnerability to climate change.
Prior to these studies, ecosystem models have assumed that tropical forests canopies showed no
seasonality (Sitch et al. 2003, Cox et al. 2004), in part, because seasonal remote sensing in the
tropics is undermined by poor atmospheric conditions (Kobayashi & Dye, 2005). Other models
represented browning vegetation and decreased activity in periods of drought (dry seasonal
periods and interannual droughts, such as ENSO; Tian et al., 1998; Botta et al., 2002).
Various hypotheses for forest response to climatic variability include:
1.
Light limitation hypothesis, in which tropical forest growth is primarily limited by
availability of photosynthetically-active radiation (PAR), with access to deep soil water
via deep tree roots minimizing oft-modeled water stresses.
2.
Diffuse radiation enhancement of Light-Use Efficiency (LUE), which is the amount of
photosynthesis that takes place per amount of PAR absorbed by the forest canopy.
3.
Leaf growth phenology (perhaps evolutionarily cued by the mechanisms of hypotheses 1
and 2)
These hypotheses for forest response to climatic variations are subject to test and investigation
by remote sensing observations, especially if mechanistically linked to ground-based studies.
However, issues remain in the interpretation and understanding of remote sensing data.
Different coarse resolution satellite products collected at high temporal frequencies from the
AVHRR and MODIS platforms show variable and inconsistent seasonal and interannual patterns
4
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
over tropical rainforests with some sensor products showing canopy drying in the dry season and
negative forest responses to drought, while other products show leaf flushing and greening in the
dry season and a positive response to drought (Figs. 1, 2). Large inconsistencies have been
reported in cross- satellite product comparisons (that include MODIS and AVHRR) for tropical
evergreen broadleaf forests (Garrigues et al. 2008) and inconsistencies have been reported
among MODIS products and tropical field observations for LAI-related products (Senna et al.
2005; Doughty and Goulden, 2008). At local tower site scales, Doughty and Goulden (2008)
showed in situ LAI measurements differed markedly from seasonal cycles of the MODIS LAI
product (MOD15) and MODIS NDVI (MOD13). Hutrya et al. (2007) also showed MODIS LAI
and EVI products not in phase with each other, with EVI best approximating the seasonal flux
tower signal of Gross Ecosystem Production (GEP).
Further, some argue that the observed remote sensing responses are due to artifact. Kobayashi
and Dye (2006) analyzed the GIMMS AVHRR- NDVI time series over the Amazon and found
the seasonal signal to be primarily dominated by cloud and aerosol contamination. Ganguly et al
(2009), for example, suggest
Figure 1. Seasonal variations in MODIS derived GPP (Pg), fpar, and that apparent tropical forest
greening in response to the
EVI products with tower Pg in SE Asia tropical forests (MaeKlong
intense Amazon drought of
Watershed Research Station, Thailand). (Huete et al. 2007)
2005 (as reported in Saleska et
al., 2007) is in fact due to
atmospheric aerosol
contamination of the surface
reflectance, rather than a true
vegetation response.
We have previously argued
(Huete et al., 2006; Saleska et
al., 2007) that by choosing
satellite products that correlate
with measurements on the ground (e.g. satellite EVI with ecosystem scale photosynthetic fluxes
measured from towers), and by a combination of appropriate selection of high quality data from
relatively uncontaminated pixels and correction for residual contamination, we arrive at
observations that are robust to the problems cited above (for example, because MODIS EVI uses
atmospherically-corrected surface reflectances and an aerosol resistance term, we believe it is
minimally affected by residual aerosols up to optical depths of over 1. However, this argument
remains to be definitively tested.
We propose a study to decisively test competing hypotheses about mechanisms underlying
green-up, including the null hypothesis that reported observations are the result of
contamination artifacts.
1.2 Example: 2005 Drought
In 2005, large areas of the southwestern Amazon basin experienced one of the most intense
droughts of the last 100 years (Marengo et al. 2008). During the 2005 drought, satellite
observations showed increases in vegetation indices, suggesting increased forest photosynthetic
5
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
capacity (Saleska et al., 2007)1, while Phillips et al. (2009), using field measurements, showed a
decline in the rate of biomass accumulation, slowing or reversing previous trends, primarily due
to an increase in tree mortality (Phillips et al. 2009). It is not currently understood whether these
seemingly divergent responses (short term increase in photosynthetic capacity during the peak
months of drought, encompassed by excess mortality integrated over several years) may in fact
be reconciled by
accounting for the differing Figure 2. Seasonal patterns of tropical forest biologic activity at
time scales and the effect
Tapajos National Forest, Brazil using indices computed from
of time lags. The divergent Hyperion hyperspectral data, including the 1st derivative of red edge
(top left), normalized difference water index (NDWI, top right) and
responses, however,
enhanced vegetation index (EVI, bottom right). The primary forest
sharply pose specific
show increasing canopy water content and greenness while the
questions which will be
addressed by our proposed pasture shows browning and drying, and the regenerating forests
show a mixed response of browning followed by greening (due to
study, including:
unique phenologies of multi-functional, herbaceous and tree layers).
• What is the relative
role of light vs. water
limitation in
controlling vegetation
response to drought?
• What are the
principal causes for
variation in forest
responses to the same
drought event (while
most areas greened
up, distinct “browndown” regions were
also observed – see
Figure 3 below).
• How does the drought
response interact with
the background pattern of seasonal dynamics? (a better understanding of the spatial
distribution and variability of seasonality in the Amazon is needed, since Amazon
ecosystems and climate are certainly not uniform and the resilience of a forest may depend
on the interactions of drought timing and length, drought strength, and forest ecosystem
phenology timing)
• Fundamentally, assuming Amazon systems have evolved to optimize resource use (lightlimitations, poor nutrient status), what happens with disturbance and climate change?
Figure 3. Spatial variability in annual Amazon rainfall (1998-2006 mean mm month-1from TRMM
satellite data, ranging from <100 in orange areas, to >275 in dark blue areas) and tower site locations
for this study (left). Example of profiles of vegetation seasonal variability with open symbols depicting
1 1
Saleska et al. (2007) reported results based on analysis of MODIS data, collection 4, the latest version then
available. Reanalysis using collection 5, the most recent revision to become available since publication, shows a
more modest green-up in the drought region. Drought area greening, however, remains statistically anomalous
under collection 5, relative to a null expectation of no difference in the amount of greening vs. browning.
6
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
average dry season periods with <60 mm monthly rainfall (right). The shaded area (July-Sept) depicts
the time period of the 2005 drought anomaly.
More fully validated observations are needed to build confidence in remotely sensed ecosystem
function in tropical forests before community acceptance of observed MODIS satellite-based
results, and to realize the full potential of the future array of space-based spectral sensors,
including the VIIRS mission and hyperspectral HyspIRI mission proposed as part of the NRC’s
Decadal Survey. Such satellite data sets are invaluable for monitoring forests and as input into
land-surface models, however, such data must be
of high quality and represent unbiased seasonal measures of Amazon ecosystem processes to
ensure more accurate future projections of Amazon response to climate change.
1.3 Objectives
We propose a study to decisively test competing hypotheses about mechanisms underlying forest
green-up during dry periods: increased LUE with more diffuse radiation, leaf-scale phenology
driven by biological rhythms, as well as the aerosol artifact hypothesis. We aim to fully
characterize tropical forest seasonality (leaf flushing, browning, litterfall) across a range of
ecosystems with varying radiation and moisture environments; understand the mechanisms of
greening (phenology, chlorophyll, LAI) and their relationships to photosynthesis (gross
ecosystem productivity, GEP) at the leaf- and canopy level; and investigate the environmental
drivers of satellite observed seasonality, including atmosphere radiation, aerosols, light and
moisture controls. Our goal is to advance and improve our understanding of seasonal biologic
activity in tropical rainforests by linking tower GEP fluxes, species-specific leaf-scale
phenology, ecosystem-scale hyperspectral observations from tower, and detailed characterization
of the local radiation regime and satellite observations.
This work will give new insight into the fundamental question of tropical forest “vulnerability to
global climate change” (Subelement 1 of the ROSES call); will produce a fusion of hyperspectral
with high-frequency moderate resolution datasets, thereby providing a focused example of how
HyspIRI Decadal Survey mission data may be used “to understand seasonal expressions and
cycles” (Subelement 2); and will synthesize and integrate substantial amounts of data from LBA,
a “past NASA Terrestrial Ecology research project” (Subelement 3).
7
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
Key Science and Remote Sensing Questions:
1. What are the controls, key environmental determinants, and interactions of biome
seasonality? (precipitation, temperature, light).
2. What are the drivers of change/ shifts in seasonality? (stress agents, multiple stressors).
3. What is the vegetation canopy signal, decoupled from aerosol and cloud seasonal signals?
4.
What are the hyperspectral signals and indices relevant for photosynthesis and stress studies.
5.
How do we scale leaf optics to canopy optics: reflectance, physical and physiological leaf
measurements before, during and after the dry season and try to scale up to leaf phenological
dynamics in the Amazon.
6.
How do we aggregate leaf phenology to canopy and landscape seasonality in multi-species/
multifunctional environments (multi-story canopies).
2. Technical approach and methodology
This study will have three components. FIRST, we will compare independent measures of the
seasonality of: (1) vegetation photosynthetic capacity (at both individual leaf and ecosystem
scales), (2) atmospheric characteristics (aerosol optical depth and cloud cover), (3) surface
radiation components (including direct and diffuse radiation), and (4) vegetation spectral
reflectances (and indices) from satellites and from in-situ AERONET sunphotometers, at four
sites across the Amazon (Santarem Forest and converted farmland; Ji-Parana Forest; and Sao
Paulo state savanna). The sites are chosen to span a range of tropical or subtropical vegetation
(from forest, to converted agricultural land, to natural savanna) and atmospheric characteristics
(from moderate to highly seasonal aerosol optical depth). SECOND, we will use hyperspectral
observations at both the surface (via continuously operating tower-mounted hyperspectral
imaging cameras) and from space (via Hyperion imagery) to diagnose key physiological
processes and further test the accuracy and interpretation of the high-frequency moderateresolution MODIS products. The resulting Hyperion/MODIS fusion will provide a model for the
optimal fusion of relatively infrequent data from the proposed HyspIRI Decadal Survey Mission
with higher-frequency data from the future VIIRS mission. THIRD, we will use the first
airborne LIDAR dataset obtained for the Amazon to parameterize a sophisticated 3-D canopy
photosynthesis model (FLiES) in order to scale up our integrated understanding of vegetation
characteristics (including leaf spectral reflectance), radiation components (including aerosol-,
cloud-, and subcanopy-influenced effects of diffuse radiation fraction and angular distribution),
and the seasonality thereof.
This work will be divided into the following tasks:
2.1 Task 1. Assessment of the seasonal expressions and cycles of Amazon tropical forest
ecosystems
In this task, we assess and characterize Amazon tropical forest green-up and brown-down
seasonality at the ecosystem/ tower and species/ leaf level scales. We will measure seasonal
expression of canopy function at multiple scales, including,
• measurements of canopy-level seasonal reflectances and indices from tower-mounted
hyperspectral cameras and with periodic top-of-canopy spectroradiometer measurements
(ASD-FR) from the tower (all 4 sites).
• pixel-based reflectances and indices from MODIS and Hyperion, derived from in-situ
atmosphere measurements and corrections from nearby AERNOET sunphotometer sites.
8
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
•
•
•
•
We will use available validation datasets from the AERONET-based Surface Reflectance
Validation Network (ASRVN, Wang et al., 2009) to obtain independent atmosphere
corrected assessments of seasonal greening. The Belterra and Abracos Hill sites are close
to the Tapajos and Jaru tower sites of this study.
tower-based flux seasonal measurements of photosynthesis (GEP). (all 4 sites)
leaf-level seasonal-spectral measurements with ASD-FR and hyperspectral digital camera
(at intensive site, km 67)
species and leaf level based gas exchange measurements of photosynthesis (at intensive
site, km 67), including local observational inventories of proportion of newly flushed vs.
old leaves, and status of colonization by epiphylls.
Seasonality of litterfall (via litter collection baskets, as in Rice et al., 2004) and LAI
observations via LAI-2000 and hemispheric photography (at intensive site, km67).
Our goal is to determine what the nominal seasonal cycles and variations of biologic activity &
metabolism are in tropical rainforests. In addition, we wish to more thoroughly characterize the
mechanisms responsible for greening, and assess the roles of 'leaf phenology', chlorophyll status,
LAI, and leaf physiology on canopy photosynthesis. Leaf flushing and green-up involve several
simultaneously changing variables, such as leaf phenology (young leaves vs older leaves vs
epiphyll-covered leaves), chlorophyll levels, leaf area, and litterfall. As Doughty and Goulden
(2008) report, one needs to consider the aggregate response of all factors to model
photosynthesis and dry-season greening in tropical forests. The approximate correspondence
between tower measures of GEP and MODIS EVI may be due to the aggregate ‘greening’ signal
provided by EVI (Figure 4).
Tropical forest seasonality is an aggregate mixture of diverse species and diverse functional
classes, which exhibit partially synchronous or asynchronous leaf flushing and flowering patterns
across species. The process of leaf flushing may typically only involve a fraction of tree species
(30-50%) and near the equator there may be two periods of flushing synchronized with solar
insolation patterns (Borchert et al. 20XX). In the more open tropical forests, and in drier
tropical forests and regenerating forests, simultaneous drying and greening associated with
understory and overstory layers and with mixed evergreen and deciduous tree species occur (Fig.
2). In this task, we aim to determine how these factors interact across diverse species to initiate
an aggregate greening response in the dry-season, as well as brown-down signal in the wet
season.
Figure 4. Combined multiple site annual averaged relationships between tower GEP and satellite EVI at
3 Monsoon Asia tropical forests. The shaded lines refer to equivalent relationships encountered in the
Amazon (rainforest + pasture) and Harvard Forest (temperate forest) sites.
2.2 Task 2. Analysis of the environmental drivers of forest seasonality
Ecological and environmental conditions across the Amazon basin are not uniform. The length
and intensity of the dry season varies widely from perhumid wet rainforests to seasonally dry
rainforests to Cerrado. Moisture regimes vary in amount and seasonality of rainfall and in soil
properties. Light regimes vary in the effects of seasonal sun-earth geometry (e.g. dual annual
9
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
peaks in solar irradiance in the equatorial zone) and in the amount, seasonality and radiative
impact of clouds and aerosols. This task will examine seasonal variations in light, moisture, and
temperature and evaluate their significance for vegetation seasonality at three principal Amazon
field sites with contrasting ecological and environmental conditions. The analysis will
characterize:
• The diurnal and seasonal variations in global, diffuse and direct fluxes of PAR, and in the
angular distribution of PAR across
the sky hemisphere.
• The relative significance of clouds
versus aerosols in controlling
seasonal variation in total and
diffuse radiation and associated
effects of diffuse conditions on
canopy photosynthesis and light use
efficiency (LUE). LUE of
photosynthesis is thought to
dynamically adapt to environmental
factors, which lead to complex
spatio-temporal variations of
photosynthesis on various scales from the leaf to the canopy level (Rascher and Pieruschka,
2008). The need to scale leaf-level physiology to ecosystem responses and climate
feedbacks has been emphasized recently in the context of global climate change research.
• The effect of observed variation in radiation, temperature, moisture conditions and tree
physiology (light response curves) on forest photosynthesis determined from observations
(CO2 eddy flux measurements) and advanced 3-D ecosystem process modeling (FLiES
model).
• The effects of changes land use, climate (radiation, precipitation), and disturbance
(drought, fire) on forest seasonality and underlying physiological controls on water and
carbon exchanges.
• The response of forest photosynthesis under light-limited conditions to variations/changes
in light conditions, including enhanced light availability associated with drought anomalies
of various durations, and moisture interactions with light-related controls.
• The role and relative significance of radiation forcings from clouds, which are seasonally
and spatially associated with rainfall, versus forcings from aerosols which are associated
with dry periods and with land disturbance (Fig. 5).
• The significance of soil type/properties (which can be spatially unique modifiers of
precipitation patterns) for forest seasonality.
Information on the seasonality of atmospheric conditions (clouds, aerosols) and solar radiation
(amount, diffuse fraction, and angular distribution of PAR) acquired in this task will enable, in
combination with results from Task 1, an unequivocal determination of the validity of satellitebased characterizations of tropical forest seasonality by decoupling of the forest and atmospheric
signals.
Figure 5. Amazon variability in spatial patterns of various drivers, including the patterns of rainfall,
aerosol, and clouds on MODIS EVI greening.
10
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
2.3 Task 3. Scaling seasonal leaf phenology to tree species to tower level canopy and to
satellite-based landscape seasonality
Tropical forests are composed of a diverse mixture of individual plant and tree species and
functional classes with unique biophysical characteristics (e.g. spectral reflectance,
photosynthetic capacity) and phenological behavior. Differences in leaf age both within and
among species may also be associated with differences biophysical characteristics. These
characteristics/behaviors may vary spatially within the canopy (vertically and horizontally) and
temporally through the seasons. The composite assemblage of individual species and leaf ages
within the forest will thus determine the integrated patterns of spectral reflectance and
seasonality observed by satellite sensors such as Landsat, Hyperion, AVHRR and MODIS, and
patterns of ecosystem-atmosphere carbon exchange to which CO2 eddy flux measurements at
tower sites are sensitive. Task 3 is designed to disaggregate these factors and examine their
relative significance determining 1) the landscape-scale patterns of seasonality and
photosynthesis exhibited by Amazon tropical forests at our four contrasting study sites, and 2)
the associated patterns of spectral
reflectance or radiance that would be
observed by current and planned
spaceborne image sensors
(multispectral, hyperspectral) in the
absence of atmospheric contamination
by clouds and aerosols. This task will
be performed by combining in situ
observations of key forest properties
measured at the leaf, tree and tree stand
levels with an advanced 3-D model of canopy radiative transfer and photosynthesis, the Forest
Light Environmental Simulator (FLiES, Kobayashi and Iwabuchi, 2008). For this task, we will
conduct the following,
• Field measurements of seasonal forest optical properties at the leaf- level and tower scale,
as described in Task 1, for a spatially representative set of canopy and understory species,
leaf age classes, and for exposed/sunlit and interior/shaded leaves.
• Leaf physiology measurements (Rubisco-limited and electron transport-limited
photosynthetic capacity, i.e. Vmax and Jmax, and stomatal resistance) at intensive site
km67.
• Existing airborne LiDAR measurements will be analyzed to characterize forest structure.
Structure will be quantified as the 3-D leaf area distribution (LAD), leaf area index (LAI),
and 3-D distribution of major non-photosynthetic, supportive elements (boles, large
branches) within the canopy and understory space (discretized as voxels) (Stoker et al.,
2009) (Spatial variations in leaf area density have greater value than landscape average
LAI).
3-D Modeling of Landscape-Scale Reflectance and Photosynthesis: The FLiES model will be
parameterized by reference to the measured spectral, physiological, structural properties
described above. The model will be used to investigate:
11
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
• how changes in leaf optics (spectral reflectivity, transmissivity) and changes in LAI influence
the landscape-scale bidirectional reflectance characteristics of the forest and spectral radiances
observed by current and planned satellite sensors (MODIS/VIIRS, HyspIRI) in the absence of
significant atmospheric (cloud/aerosol) contamination,
• how changes in environmental conditions (amount and diffuse fraction/directionality of PAR,
temperature, soil moisture and VPD) affect the landscape-scale rates of photosynthesis, and
• how the observed seasonal changes in forest properties and environmental conditions interact
to determine the seasonality of forest photosynthetic activity at the landscape scale.
• Model simulations will be validated by comparisons with estimates of forest photosynthesis
and carbon exchange from the CO2 eddy flux measurements at the tower sites.
HyspIRI- MODIS/VIIRS Data Fusion: Although MODIS- and the future VIIRS-EVI can
potentially provide an overall seasonal correspondence with tower flux measures of GEP (Fig.
4), EVI by itself remains ambiguous to the mechanisms and causes of greening. Finer spectral
and spatial resolution satellite data from the HyspIRI sensor will be needed to fully characterize
complex tropical forest seasonality, involving diverse species assemblages and functional
classes, leaf ages, light-use efficiencies, and canopy chlorophyll and moisture status. One needs
the right combination of spectral, spatial or temporal resolutions to adequately characterize the
complex seasonal dynamics of tropical forests. High temporal frequency measurements are
critical in obtaining sufficient
acquisitions of cloud-free data and
achieve greater sensitivity to seasonal
variations, while canopy physiology and
biochemical canopy features, diverse
species assemblages, and forest
disturbance are best defined at finer
resolutions and spectral fidelity. The
combined use of hyperspectral
measurements, like HyspIRI, with finer
temporal moderate resolution data
(MODIS/VIIRS) will facilitate,
• a more complete understanding of
seasonal expressions and cycles in
tropical forest ecosystems,
including an assessment of their functional groups and key physiological processes,
• test the accuracy and interpretation of the high-frequency moderate-resolution MODIS
products,
• time series continuity and translation across sensors by convolving hyperspectral imagery
to the spectral and spatial resolutions of fine/ moderate and coarse resolution bandpasses
for identical atmosphere, viewing, and surface conditions (thereby simulating future and
past sensor systems) and allow,
• information mining from multi-source and multi-scale, spatial-temporal-spectral data for
seasonal- biochemical and structural analyses of tropical forests.
• utilize spectral signature based algorithms and mixture modeling approaches to convert
spectra to structural, biochemical, and phenologic properties of ecosystems (chlorophyll,
leaf color, leaf age) at key wet-dry seasonal transition and peak periods.
12
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
Figure 6. A seasonal set of 6 hyperspectral Hyperion images acquired through the 2001-2002 dry season
periods (Jul-Dec) over Tapajos National Forest vicinity (top left). Spectral signatures of primary forest,
pasture, and regenerating forests over Tapajos National Park in early and late dry season (right) and
their subtracted difference (bottom left).
2.4 Data Acquisition from four tower sites
We will synthesize extensive datasets from past observations, and acquire critical new
measurements from four sites across the Amazon. The four sites (see Fig. 3) are:
(1)
Santarem primary forest site (km67 tower site in the Tapajos National Forest) (Saleska et
al., 2003; Hutyra et al., 2007)
(2)
Santarem agricultural site (km77 tower site in privately owned farmland ~15 km distant
from the Tapajos forest site, Sakai et al., 2004);
(3)
Ji-Parana Forest site (Reserva Jaru eddy flux tower site, von Randow et al., 2004); and
(4)
Sao Paulo state savanna (tower site in Pe-de-Gigante reserve).
These sites were among those established in the 1999-2002 time period by the ecological
component of the Large Scale Biosphere-Atmosphere experiment in Amazonia (LBA) (Keller et
al., 2004). The sites are chosen from those currently in operation to span a range of tropical or
subtropical vegetation: from forested sites (near Santarem and Ji-Parana), to converted
agricultural land (near Santarem), to natural savanna (subtropical site in Pe-de-Gigante reserve).
They also span a range in atmospheric characteristics from moderately seasonal aerosol optical
depth (at Tapajos), to Ji-Parana, typically at the center of the thickest part of the seasonal aerosol
plume in the southern Amazon basin (Oliveira et al, 2007). Also there are AERONET sites
located near the Tapajos National Forest (Belterra) and Jaru Forest Reserve (Abracos Hill).
2.5 Measurements
2.5.1 Ongoing Automated tower measurements. Eddy covariance fluxes of sensible heat, CO2,
and H2O, are currently made at these sites, as are basic radiation measurements (PAR, net
radiation, and short and long-wave radiation components). AERONET CIMEL sun-photometers
(giving information on aerosol optical depth) have been in operation at two sites near three of the
towers: the Abracos Hill site, at the Fazenda Nossa Senhora Aparecida (10◦45’S and 62◦22’W),
about 100 km from the Reserva Jaru flux tower, and at Belterra (02◦39’S and 54◦57’W) close to
both the km67 forest and km77 agricultural tower sites. The km67 eddy flux tower also includes
continuous measurements of total and diffuse radiation (SPN1 Sunshine Pyranometer, Delta-T
Devices, Ltd.). For each site we will estimate ecosystem-scale photosynthesis from Gross
Ecosystem Productivity (GEP) of CO2, an estimate of the magnitude of ecosystem-scale
photosynthesis, as the difference between NEE and ecosystem respiration (Reco), where Reco is
assumed equal to nighttime NEE (Saleska et al. 2003, Goulden et al. 2004, Hutyra et al. 2007).
We will estimate ecosystem scale photosynthetic capacity, Pc, (an index that is largely
13
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
independent of varying light levels), as the mean GEP for a fixed range of PAR (725 > PAR <
925 mol m-2 s-1), as defined in Hutyra et al. (2007).
2.5.2 Proposed new measurements for this project. We propose to supplement ongoing
measurements at each of these sites with new measurements of:
• Automated hyperspectral imaging systems (‘Surface Optics Corporation’) (3 cameras, one for
each of the Santarem, Ji-Parana, and Savanna sites). These will be mounted at tower top and
logged continuously, providing high resolution hyperspectral images at each site, giving leaf
reflectance measurements at the surface at each site. The camera near Santarem will be
exchanged bi-weekly between the forest site (km67) and the agricultural site (km77).
• Total and diffuse radiation sensors (BF3 Sunshine sensor, Delta-T Devices, Ltd: two sensors,
one each for Ji-Parana and Pe-de-Gigante Savanna sites). These will be mounted on towers at
these two sites, to complement the existing total and diffuse Sunshine sensor (SPN1, Delta-T)
already installed at the Santarem km67 forest site.
• High dynamic range radiometric all-sky (hemispheric) imaging systems (based on CMOS
image sensors; two instruments, one each for the Santarem and Ji-Parana sites) will measure
sky radiance distributions and cloud/aerosol dynamics.
2.5.3 Additional measurements at intensive core site (Santarem km67 forest site). We propose
additional species specific leaf-scale measurements at one site chosen for additional intensive
measurements needed to fulfill project goals. These include new leaf-scale measurements of
photosynthetic capacity and hyperspectral reflectance, and use of recently acquired ecosystemscale airborne LIDAR measurements.
Ongoing ecological measurements include: (1) Live Tree dynamics (diameter increment,
mortality and recruitment of trees), based on annual re-surveys (past surveys were in 1999, 2001,
2003, 2005, 2007, 2008, and 2009, giving long-term fluxes from tree growth, recruitment, and
mortality, using methods detailed in Rice et al., (2004), and using a ground-based portable
canopy Lidar (PCL) system; (2) Litterfall and LAI seasonality (via litter baskets, LAI-2000, and
hemispheric photos).
New leaf-scale measurements: Species-specific measurements of in-situ leaf photosynthetic
capacity (via Licor-6400 portable leaf-gas exchange system), paired with leaf hyperspectral
absorbance measurements (via ASD camera system). Measurements will be made from a
permanent canopy access tower, with a broader spatial sample obtained from a canopy access net
temporarily installed in sample tree canopies by professional tree climbers. Measurements will
be made on target canopy trees at least 4 times per year, both at the canopy top and in the
subcanopy (to obtain measurements on both sun and shade leaves).
2.5.4. Recently acquired airborne LIDAR data: Aircraft LIDAR surveys were flown over the
Brazilian Amazon in June 2008 with support from the U.S. National Science Foundation (NSF
grant #0721140, PI: Saleska, University of Arizona), via a contract to Esteio Engenharia e
Aerolevantamentos, a Brazilian commercial survey company with over 30 years of experience in
aircraft surveying and mapping. The sensor was a Leica Lidar (the Leica ALS50-II Airborne
LiDAR Sensor), mounted on board a Navajo EMB 820C aircraft. Surveys were conducted over
the Tapajos National Forest near Santarem (including a 400 ha area encompassing the km67
eddy tower), at an altitude of 800m, with the sensor operating at 120 kHz. Nominal pulse
density on the ground was 9 points/m2 (with a maximum scan angle of 10 degrees). Use of
14
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
2.6. FLiES model (3-d canopy model)
The interactive effects of forest canopy properties (structural, optical, physiological) and
atmospheric radiation conditions (amount, angular distribution, spectral composition) on 3-D
canopy radiative transfer and photosynthesis will be investigated with the Forest Light
Environmental Simulator (FLiES) (Iwabuchi and Kobayashi, 2008) (Fig. 7). The FliES model is
based on the Monte Carlo ray tracing method, and employs a leaf photosynthesis model
(Farquhar et al., 1980) to calculate canopy level photosynthesis from specified canopy
parameters (3-d structure, leaf/stem optical properties) and environmental variables (irradiance,
temperature, VPD). The 3-D canopy structure at the tower sites will be quantified through
analysis of existing aircraft-based LiDAR measurements. The strategy for radiation measurement
is described in a subsequent section.
Figure 7. FLiES model simulations of absorbed photosynthetically active radiation (APAR) for tropical
forest canopy under broken clouds with high diffuse fraction (left) and clear sky conditions (right) for
the same solar zenith angle. 2-D variation in APAR for sample vertical slice through canopy is
displayed.
15
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
The FLiES model will be used to evaluate of the relations between observed variations in
atmospheric conditions (global and diffuse PAR, angular PAR distribution) and forest canopy
photosynthesis simulated with the FLiES model and validated against CO2 exchange rates
estimated by the eddy covariance technique at the tower sites. This analysis will enable
identification of the atmospheric/radiation conditions associated with maximum and/or optimal
rates of canopy photosynthesis and their seasonal timing vis a vis the in situ and satellite-based
observations of canopy properties. Moreover, we expect to detect the conditions (PAR,
precipitation/soil moisture) and seasonal timing when canopy photosynthesis enters or emerges
from primarly light-limited or water-limited conditions, if such a shift occurs.
Figure 8. Sample image (RGB composite) of sky conditions at a tropical site from a prototype high
dynamic range sky imager (right). The image was captured at the instance indicated by the arrow in the
plot of diurnal PPFD (left). Angular distribution of PAR irradiance is estimated from pixel radiances
measured by red, green and blue channels of the CMOS image sensor. The system employs an innovative
high dynamic range imaging technique that enables measurement of the full range of observed sky
radiances (including the solar disk) without saturation.
2.7 Seasonal patterns of solar radiation.
We will monitor global and diffuse fluxes of PAR and the angular distribution of PAR, and
characterize their variability on diurnal, seasonal, and interannual time scales. Global and
diffuse fluxes will be measured with existing solar radiation sensors (SPN1 Sunshine
Pyranometer, Delta-T Devices, Ltd.). The direct flux will be calculated as the difference of the
global and diffuse components. The PAR angular distribution will be measured with a CMOS
camera-based system for radiometric sky imaging. The instrument employs an innovative, highdynamic range imaging technique that enables measurement of the full range of sky radiances
(including the solar disk) without radiometric saturation and without a shading device (Fig. 8).
The PAR angular information is critical to achieve realistic and accurate simulations of 3-D
can
opy
radi
ativ
e
tran
sfer
and
pho
tosy
nth
esis
with the FLiES model.
3. Management plan
3.1 Management structure and collaborations
The project will be managed by Dr. Huete (P.I.) at University of Arizona. Dr. Huete, in
cooperation with Dr. Saleska (also at Univ. Arizona) will advise the post-doctoral researcher and
one graduate student researcher at Univ. Arizona. Dr. Huete will maintain communication and
coordination among the project team members and their respective activities. Collaborations
16
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
outside of Univ. Arizona that are integral to the project are with Dr. Dye (Co-I at USGS) for
environmental observations and analysis; Dr. Kobayashi (Collaborator at UC Berkeley and
JAMSTEC) for 3-D canopy modeling; and with Brazilian team members (Co-I’s Dr.
Shimabukuro, Dr. da Rocha and Dr. Manzi) for field work, tower instrumentation, and sharing of
tower flux measurements of the four sites.
Drs. Huete and Saleska will provide primary leadership for executing the assessment of the
Amazon tropical forest seasonality and its scaling from leaf to landscape levels (Task 1), with
support from the post-doctoral researcher and one graduate student. Dr. Dye, in collaboration
with Drs. Saleska and Huete, will direct observation and analysis of the environmental drivers of
forest seasonality (Task 2), with Dye’s primary focus on radiation-related factors. A second
graduate student at Univ. Arizona will support this task, and will be co-advised by Drs. Dye and
Huete. Dr. Kobayashi is a research scientist at JAMSTEC in Japan and is currently a visiting
scientist at UC Berkeley in Dr. Baldocchi’s lab. He will perform modeling and analysis of 3-D
canopy radiative transfer and photosynthesis with the FLiES model, with close interaction with
the Postdoc and 1 graduate student for model parameterization with results from field
observations. Our Brazilian investigators are vital to this work effort. Drs. da Rocha and Manzi
will collaborate with us in the field, help maintain our instrumentation, and share with us tower
flux and other measurements as described in Table 1. Dr. Shimabukuro, from the Brazilian
Space Agency has extensive experience with MODIS, Hyperion, and Landsat imagery and will
help with logistical support for phenology measurements, including the field participation of
graduate students from INPE.
3.2 Milestones and timeline
Year 1. (1) Set-up of instrumentation (atmosphere and canopy spectral cameras); (2)
Comparisons of independent measures of seasonality, both in-situ (leaf and tower-based optical
and gas exchange measurements) and from satellite data (Hyperion and MODIS) for two of the
study sites; (3) Calibration of the hyperspectral imaging cameras; (4) Analysis of seasonal
atmospheric characteristics (aerosol optical depth and cloud cover), and seasonality in surface
radiation components (including direct and diffuse radiation); and (5) We will also begin initial
tests and parameterization of the FLIES model with leaf optical measurements and the airborne
LIDAR data.
Year 2. (1) Set-up instrumentation at the remaining 2 sites; (2) analyze the AERONET
(ASVRN) retrieved surface reflectances at the pixel level and compare with standard MODIS
products and Hyperion imagery; (3) initiate scaling and aggregation studies from leaf to canopy
scale; (4) conduct diffuse/ direct radiation analyses and the influence (direct and indirect) of
aerosols on observed greening and LUE; (5) investigate the use of key diagnostic species to
characterize canopy seasonal greening and browning.
Year 3. (1) We expect to complete all tasks outlines with complete scaling up assessments from
leaf to canopy to satellite-based spectral measures and their coupling with vegetation
physiological conditions and tower-based GEP measurements; (2) attempt some downscaling
(tower to species and leaf level) or model inversion studies; (3) conduct sensitivity component
analyses of relative signal contributions from the canopy and the atmosphere in satellite datasets;
(4) complete the assessment of the relative contributions of species phenology, leaf area, and
chlorophyll content on observed seasonal greening and GEP measurements.
3.3 Expected outcomes and products
17
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
• Improved knowledge of seasonal expressions of canopy biophysical and chemical properties
and processes (photosynthesis, chlorophyll, LAI, leaf color, FPAR, LUE)
• A parameterized and sophisticated 3-D canopy photosynthesis model (FLiES) in order to scale
up tropical forest characteristics (including leaf spectral reflectance), radiation components
(including aerosol-, cloud-, and subcanopy-influenced effects of diffuse radiation fraction and
angular distribution), and the seasonality thereof.
• The merger and potential fusion of hyperspectral and moderate resolution data in space and
time for a more complete characterization of tropical forest greening, browning, and
seasonality of key physiological processes (Hyperion-MODIS; HyspiIRI- MODIS/VIIRS).
• An improved understanding of Amazon Basin responses and/or resilience to drought intensity,
timing, and length.
3.4 Facilities
The Terrestrial Biophysics and Remote Sensing (TBRS) Lab at the University of Arizona
maintains computing laboratory resources dedicated to remote sensing data processing,
analysis, algorithm development, and modeling. The main computing system consists of PCLinux cluster and web server and PC workstations integrated with an SGI-UNIX workstation
for image and field data storage/sharing among different platforms. ENVI image processing
(with IDL) and ArcGIS software packages are maintained within the science computing
facility. Additional equipment such as 2 portable ASD-FR spectroradiometers, LAI-2000,
integrating sphere, sun-photometer, and quantum sensors are used in the field for validation
and research purposes.
4. References
Betts RA, Cox PM, Collins M, Harris PP, Huntingford C, Jones CD, The role of ecosystem-atmosphere
interactions in simulated Amazonian precipitation decrease and forest dieback under global climate
warming, THEORETICAL AND APPLIED CLIMATOLOGY 78 (1-3): 157-175 JUN 2004.
Borchert R, Rivera G, Hagnauer W (2002) Biotropica 34:27–39.
Botta, A., N. Ramankutty and J. A. Foley (2002), Long-term variations of climate and carbon fluxes over
the Amazon basin. Geophysical Research Letters, 29, Article-1319.
Cox, P.M., R. A. Betts, C. D. Jones, S. A. Spall, I. J. Totterdell (2000), Acceleration of global warming
due to carbon-cycle feedbacks in a coupled climate model. Nature 408 (6809): 184-187.
Cox, P. M., P. P. Harris, C. Huntingford, R. A. Betts, M. Collins, C. D. Jones, T. E. Jupp, J. A. Marengo,
and C. A. Nobre. 2008. Increasing risk of Amazonian drought due to decreasing aerosol pollution. Nature
453:212-216.
Doughty, C. E., and M. L. Goulden (2008), Seasonal patterns of tropical forest leaf area index and CO2
exchange, J. Geophys. Res., 113, G00B06, doi:10.1029/2007JG000590.
Farquhar, G.D., von Caemmerer, S., and Berry, J.A. (1980). A biochemical model of photosynthetic CO2
assimilation in leaves of C3 species. Planta, 149, 78−90.
Ganguly, S., A. Samanta, H. Hashimoto, Y. Knyazikhin, R.R. Nemani, R.B. Myneni, (2009), Amazon
Forests did not green-up during 2005 drought. Manuscript (pers. comm..).
18
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
Garrigues, S, R Lacaze, F Baret, JT Morisette, M Weiss, JE Nickeson, E Fernandes, S Plummer, NV
Shabanov, RB Myneni, Y Knyazikhin, and W Yang. 2008. Validation and intercomparison of global Leaf
Area Index products derived from remote sensing data. Journal of Geophysical Research
113:doi:10.1029/2007JG000635.
Huete, A. R., K. Didan, Y. E. Shimabukuro, P. Ratana, S. R. Saleska, L. R. Hutyra, W. Yang, R. R.
Nemani, and R. B. Myneni. 2006. Amazon rainforests green-up with sunlight in dry season. Geophysical
Research Letters 33:L06405, doi:06410.01029/02005GL025583.
Huete, AR, N Restrepo-Coupe, P Ratana, K Didan, SR Saleska, K Ichii, S Panuthai, and M Gamo. 2008.
Multiple site tower flux and remote sensing comparisons of tropical forest dynamics in Monsoon Asia.
Agricultural and Forest Meteorology.
Hutyra, L. R.. J. W. Munger, S. R. Saleska, E. Gottlieb, B. C. Daube, A. L. Dunn, D. F. Amaral, P. B. de
Camargo, S. C. Wofsy (2007), Seasonal controls on the exchange of carbon and water in an Amazonian
rainforest, JGR-Biogeosci, 112, G03008, doi:10.1029/2006JG000365.
Keller, M., Alencar, A., Asner, G.P., Braswell, B., Bustamante, M., Davidson, E., Feldpausch, T.,
Fernandes, E., Goulden, M., Kabat, P., Kruijt, B., Luizao, F., Miller, S., Markewitz, D., Nobre, A.D.,
Nobre, CA., Filho, N.P., da Rocha, H., Dias, P.S., von Randow, C., Vourlitis, G.L. (2004). Ecological
research in the large-scale biosphere-atmosphere experiment in Amazonia: early results. Ecological
Applications 14(4): S3-S16.
Kobayashi, H., and Dye, D.G. (2005) Atmospheric conditions for monitoring the long-term vegetation
dynamics in the Amazon using normalized difference vegetation index. Remote Sensing of Environment,
97(4), 519-525.
Kobayashi, H., and Iwabuchi, H. (2008) A coupled 1-D atmosphere and 3-D canopy radiative transfer
model for canopy reflectance, light environment, and photosynthesis simulation in a heterogeneous
landscape. Remote Sensing of Environment, 112, 173-185.
Malhi, Y., L. E. O. C. Aragao, D. Galbraith, C. Huntingford, R. Fisher, P. Zelazowskia, S. Sitch, C.
McSweeney and P. Meir (2009), Exploring the likelihood and mechanism of a climate-change induced
dieback of the Amazon rainforest. Proc. Nat. Acad. Sci. early edition
(www.pnas.org/cgi/doi/10.1073/pnas.0804619106)
Marengo, J.A., C.A. Nobre, J. Tomasella, M.D. Oyama, G.S. de Oliveira, R. de Oliveira, H. Camargo,
L.M. Alves, I.F. Brown. (2008). The drought of Amazonia in 2005. J. of Climate, 21: 495-516.
Myneni, R. B., W. Yang, R. R. Nemani, A. R. Huete, R. E. Dickinson, Y. Knyazikhin, K. Didan, R. Fu,
R. I. Negron Juarez, S. S. Saatchi, H. Hashimoto, N. V. Shabanov, B. Tan, P. Ratana, J. L. Privette, J. T.
Morisette, E. F. Vermote, D. P. Roy, R. E. Wolfe, M. A. Fiedl, S. W. Running, P. Votava, N. El-Saleous,
S. Devadiga, Y. Su, and V. V. Salomonson. 2007. Large seasonal swings in leaf area of Amazon
rainforests. Proceedings of the National Academy of Science:doi:10.1073/pnas.0611338104.
Oliveira, RS, L Bezerra, EA Davidson, F Pinto, CA 1 Klink, D Nepstad, and A Moreira. 2005. Deep root
function in soil water dynamics in cerrado savannas of central Brazil. Functional Ecology 19:574-581.
Phillips, OL, LEOC Aragão, SL Lewis, JB Fisher, J Lloyd, G Lopez-Gonzalez, Y Malhi, A Monteagudo,
J Peacock, CA Quesada, G van der Heijden, et al., 2009. Drought sensitivity of the Amazon rainforest.
Science 323:1344-1347.
19
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
Rascher U. and Pieruschka R., 2008, Spatio-temporal variations of photosynthesis ¬ The potential of
optical remote sensing to better understand and scale light use efficiency and stresses of plant
ecosystems. Precision Agriculture, 9, 355-366.
Sakai, R. K., D. R. Fitzjarrald, O. L. L. Moraes, R. M. Staebler, O. C. Acevedo, M. J. Czikowsky, R. da
Silva, E. Brait and V. Miranda (2004), Land-use change effects on local energy, water and carbon
balances in an Amazonian agricultural field. Glob. Change Biol. 10(5): 895-907.
Saleska, S. R., K. Didan, A. R. Huete, and H. R. da Rocha. 2007. Amazon forests green up during 2005
drought. Science:DOI: 10.1126/science.1146663.
Saleska, S. R., S. D. Miller, D. M. Matross, M. L. Goulden, S. C. Wofsy, H. R. da Rocha, P. B. de
Camargo, P. Crill, B. C. Daube, H. C. de Freitas, L. R. Hutyra, M. Keller, V. Kirchhoff, M. Menton, J. W.
Munger, E. H. Pyle, A. H. Rice, and H. Silva. 2003. Carbon in Amazon forests: Unexpected seasonal
fluxes and disturbance induced losses. Science 302:1554-1558.
Senna, M. C. A., M. H. Costa, and Y. E. Shimabukuro. 2005. Fraction of photosynthetically active
radiation absorbed by Amazon tropical forest: A comparison of field measurements, modeling, and
remote sensing. Journal of Geophysical Research 110:doi:10.1029/2004JG000005.
Sitch, S., C. Huntingford, N. Gedney et al. P.E. Levy, M. Lomas, S.L. Piao, R. Betts, P. Ciais, P. Cox, P.
Friedlingstein, C.D. Jones, I.C. Prentice, F.I. Woodward (2008). Evaluation of the terrestrial carbon
cycle, future plant geography and climate-carbon cycle feedbacks using five Dynamic Global Vegetation
Models (DGVMs) Glob. Change Biol. 14 (9) 2015-2039.
Stoker, J.M., 2009, Volumetric visualization of multiple-return Lidar data - using voxels, Photogramm.
Eng. & Remote Sens., 75:109-112.
Tian H., J. M. Melillo, D. W. Kicklighter, A. D. McGuire, J.V.K. Helfrich, B. Moore and C.J.
Vorosmarty (1998), Effect of interannual climate variability on carbon storage in Amazonian
Ecosystems. Nature 396: 664-667.
von Randow C, Manzi AO, Kruijt B, et al. (2004) Comparative measurements and seasonal variations in
energy and carbon exchange over forest and pasture in South West Amazonia. Theoretical and Applied
Climatology, 78, 5-26.
Wang, Y., A. Lyapustin, J. L. Privette, J. T. Morisette, B. Holben, ”Atmospheric
Correction at AERONET Locations: A New Science and Validation Data Set,” IEEE
TGARS, 2009 (in press).
BRAZIL TOWER SITE
Equipment & Facilities
Existing Facilities in Brazil from prior support from NASA, the National Science Foundation,
and Brazilian sources
Santarem tower sites
As part of the NASA-sponsored portion of the Large-Scale Biosphere Atmosphere
Experiment in Amazonia (LBA) program, international collaborations established the km 67
Primary Forest site in the Tapajós National forest, near Santarém (Harvard University and
University of Sao Paulo), and the km 77 pasture/agricultural site in converted land about 15 km
from the forest site (SUNY-Albany and the Brazilian Federal University of Santa Maria). The
sites were fully equipped at NASA expense, including diesel generator, road access, and huts for
housing instrumentation for atmospheric measurements and eddy flux observations. Custom20
Scaling photosynthesis in tropical systems: from forest to savanna, from seasons to extreme events (Huete)
designed and built eddy flux instrumentation (valued at >$1.2M with original construction and
installation costs supported by NASA), operated continuously at km 77 and km67 from
September 2000 and April, 2001, respectively. The original km 67 tower was destroyed by a
treefall in January 2006, and the complete eddy flux system (which survived the tree fall) was reinstalled in 2008 with support from the National Science Foundation through a Partnerships for
International Research and Education (PIRE) grant. Km77 flux and radiation measurements
ceased with the end of NASA LBA support, but these will be restarted in late 2009 with support
from Brazilian LBA sources (by Brazilian Collaborator da Silva).
Southern Amazon Forest and Savanna sites
An LBA forest site for eddy flux measurements was established in 1999 at the Jaru
biological reserve as part of an international collaboration between the Brazilian National
Institute for Amazonian Research (INPA), and the European Union. The site of the original Jaru
tower was occupied by native groups, and the tower was moved, and is now operated nearby by
the Brazilian LBA offices of INPA (under the direction of Brazilian Collaborator Manzi).
Collaborator da Rocha established eddy flux measurements at the savanna site at the Pe-deGigante reserve in 2004 with Brazilian support, and these measurements are ongoing.
Data from all of these sites are publicly available, at the ORNL DAAC, see
ftp://daac.ornl.gov/lba/carbon_dynamics/CD32_Brazil_Flux_Network/data/
21